Marketing ROI is a simple fraction that most businesses never actually compute: what came back, divided by what went in. It stays uncomputed not because the maths is hard — it's division — but because the inputs take honesty to assemble. This guide walks through building a real ROI picture for a small business, the traps that flatter or damn your numbers unfairly, and how to use our free ROI calculator to make the whole thing a two-minute habit instead of an annual argument.
The formula, and what goes in it
Return on investment is revenue attributable to marketing, minus the cost of that marketing, divided by the cost. The revenue side must be attributable: money you can reasonably trace to the activity — enquiries from the campaign that became invoices — not total company revenue during the campaign. The cost side must be complete: ad spend, agency or freelancer fees, tools, and the one everyone omits — your own time. Ten hours a month of the owner's attention is a real cost; a "free" channel that consumes your week isn't free. Get both sides honest and the fraction becomes trustworthy.
Lifetime value: the input that changes everything
The most consequential choice is what a customer is worth. Count only the first transaction and most marketing looks mediocre; count the realistic lifetime — repeat purchases, renewals, referrals — and the same campaign can look excellent. Neither extreme is honest by default: use the repeat behaviour you can actually observe in your own customer history, not an aspirational figure. A clinic whose average patient returns for years and a retailer whose average buyer purchases once must judge the identical cost-per-customer completely differently.
Attribution without the enterprise theatre
Big companies run multi-touch attribution models; a small business needs two humble instruments used consistently. First, ask every new customer "how did you find us?" and write the answer down — imperfect, but directionally powerful and free. Second, keep trackable paths where they're natural: campaign-specific landing pages, a distinct phone number on ads if call volume matters, UTM tags on links. Perfect attribution is unachievable and unnecessary; consistent attribution is achievable and sufficient, because you're comparing months against months measured the same way.
Time horizons: the fairness rule
Channels mature at different speeds, and judging them on one clock is how good strategies get killed. Paid ads produce data in weeks — judge them in months. SEO and content compound over quarters — judge them in half-years, and remember the asset keeps paying after you stop feeding it, which ad spend never does. The fair comparison isn't "which channel returned more this month" but "which returns more per dollar over its natural cycle." Run both clocks honestly and the answer for most small businesses is a sequence, not a winner: ads for immediate evidence and cash flow, organic for compounding position, each funding the other.
What to do with a bad number
A negative ROI is information, not a verdict. Diagnose in order: is the revenue side undercounted (missing attribution, ignoring lifetime value)? Is the leak in traffic quality (wrong keywords, wrong audience) or in conversion (a landing page that loses the visitors you paid for — run it through the free website audit)? Only after those checks does "this channel doesn't work for us" become a legitimate conclusion. And when a channel truly fails your arithmetic, kill it without sentiment — sunk cost is not a strategy, and the budget has better uses, including with someone accountable for the number.
A quarterly ROI ritual that takes thirty minutes
Once a quarter, with your actual numbers open: list every marketing cost for the period, including hours of your own time at an honest rate. List new customers and, using your "how did you find us?" answers, attribute each to a channel — "unsure" is an acceptable column; don't force false precision. Multiply customers by your observed lifetime value, divide by channel, and write the three resulting numbers where you'll see them: best channel, worst channel, and blended overall. Then make exactly one decision — feed the best, fix or kill the worst — and book the next quarter's session before closing the spreadsheet. The ritual matters more than the sophistication: a business that runs this crude version every quarter will out-decide one that plans a perfect attribution model it never builds.
One distinction saves a lot of bad decisions: blended versus marginal ROI. Blended is everything averaged — useful for the annual picture, useless for the next dollar. The next dollar should go where marginal return is highest: the campaign that would profitably spend more if fed, not the channel with the best average built on cheap early wins. Accounts routinely show a flattering blended number while their last increment of spend loses money — and the reverse, a mediocre average hiding a segment that would scale beautifully. When you review quarterly, ask both questions: how did we do overall, and where would one more dollar work hardest?
The honest summary
Attributable revenue in, complete costs out, lifetime value chosen honestly, channels judged on their natural clocks, and one consistent attribution habit. Do that quarterly and marketing stops being a faith-based line item — it becomes the most measurable spend in your business, which is exactly what it should have been all along.